A novel variable neighborhood search for the offloading and resource allocation in Mobile-Edge Computing
Mohamed younes kaci, Malika Bessedik, Amina Lammari · International Journal of Computers and Applications · 2021
New mobile applications, with increasingly demanding IT requirements on the one hand, and limited battery capacity on the other, require appropriate solutions to extend battery life while reducing latency. To cope with this problem, a new emerging concept known as ‘Multi-access Edge Computing’ (MEC) has been introduced. MEC brings computing and storage resources to the edge of the mobile network, enabling highly demanding applications to run in the mobile user equipment, using the network resources as well as possible, while minimizing latency. Offloading computations in the MEC architecture is an NP-hard problem [Jiang W, Gong Y, Cao Y, et al. Energy delay-cost tradeoff for task offloading in imbalanced edge cloud based computing. arXiv preprint arXiv:1805.02006; 2018] that has aroused the interest of several researchers, particularly over the last decade. Several algorithms have therefore been proposed to solve it. In this work, we propose a novel specific Local Search (LS) to solve the offloading problem in the case of multi-mobile users with multiple tasks per user. The proposed LS uses three neighborhood strategies and takes into account the problem specificities. Tests performed over generated instances of different sizes, as well as comparisons with works dealing with the same problem, prove the efficiency of the proposed Variable Neighborhood Search (VNS) in terms of cost and runtime, for solving the offloading problem in MEC architectures.